Typhoid Toxin Provides a Window Into Typhoid Fever and the Biology Of

Typhoid toxin provides a window into typhoid fever INAUGURAL ARTICLE and the biology of Salmonella Typhi

Jorge E. Galána,1

aDepartment of Microbial Pathogenesis, Yale University School of Medicine, New Haven, CT 06536

This contribution is part of the special series of Inaugural Articles by members of the National Academy of Sciences elected in 2012.

Contributed by Jorge E. Galán, April 25, 2016 (sent for review April 7, 2016; reviewed by Gordon Dougan and Stanley Falkow) Salmonella Typhi is the cause of typhoid fever, a disease that has studies of related Salmonella serovars such as S. Typhimurium in challenged humans throughout history and continues to be a major the mouse model of infection (12, 13). Although these studies have public health concern. Unlike infections with most other Salmonellae, revealed major insight into core-common Salmonella virulence which result in self-limiting gastroenteritis, typhoid fever is a life- traits, they have been less informative about specific aspects of threatening systemic disease. Furthermore, in contrast to most Sal- typhoid fever in humans. In particular, these studies have not monellae, which can infect a broad range of hosts, S. Typhi is a strict provided information on the pathogenesis of typhoid fever’s human pathogen. The unique features of S. Typhi pathogenesis and unique symptomatology nor have they provided insight into the its stringent host specificity have been a long-standing puzzle. The mechanisms of S. Typhi’s strict host specificity. Comparison of discovery of typhoid toxin not only has provided major insight into the genome sequences of S. Typhi with those of nontyphoidal these questions but also has offered unique opportunities to develop Salmonellae reveals relatively few unique genes, most of which novel therapeutic and prevention strategies to combat typhoid fever. are associated with lysogenic bacteriophages or genomic islands, and fewer encode potential virulence factors (15, 16, 18, 19). This Salmonella Typhi | typhoid fever | bacterial pathogenesis | bacterial article will focus on typhoid toxin, an S. Typhi virulence factor toxins | cell autonomous immunity recently discovered in my laboratory, which is also encoded by the related typhoidal serovar Salmonella Paratyphi A but it is

he bacterial pathogen Salmonella enterica serovar Typhi largely absent from nontyphoidal Salmonella enterica serovars. MICROBIOLOGY T(S. Typhi) is the cause of typhoid fever, a systemic, life- The study of typhoid toxin has revealed unique insight not only threatening disease of humans (1–5). A related but evolution- into the pathogenesis of typhoid fever but also into the molecular arily distinct Salmonella serovar, S. Paratyphi serotype A, can also bases for S. Typhi’s remarkable host specificity. cause an enteric fever type of disease virtually indistinguishable from typhoid (6, 7). These Salmonella serovars are therefore com- Discovery of Typhoid Toxin monly referred to as “typhoidal Salmonellae.” Typhoid fever is one The discovery of typhoid toxin began when our laboratory of the oldest human diseases about which we have written records in identified a toxic activity associated with S. Typhi that exhibited western literature. It is thought to have been the cause of the many of the features seen in cells intoxicated with an exotoxin “plague of Athens,” which ravaged the city of Athens during the known as “cytolethal distending toxin” (CDT) (20–22). Similar Peloponnesian war of 430 B.C. In his “History of the Pelo- to cells intoxicated with CDT, S. Typhi-infected cells seemed ponnesian War,” Thucydides, who himself suffered from the dis- distended, with nuclei double the normal size (Fig. 1). Flow ease, described the epidemic and its devastating effect in great and cytometric analysis of DNA content determined that, similar to dramatic detail (8). Although his eloquent description of the disease cells intoxicated with CDT, these cells were arrested in the G2/M was not specific enough to implicate a particular etiological agent, a phase of the cell cycle (20) (Fig. 1). Intoxication required bac- recent study of teeth recovered from an ancient Greek burial site terial infection, and no toxic activity was detected in supernatants dating back to that time detected DNA sequences suggesting that or cell lysates of bacteria grown in standard culture medium (20). S. Typhi was the actual cause of this epidemic (9). In modern times, In fact, toxicity was dependent on the ability of S. Typhi to in- typhoid fever and S. Typhi are intimately linked to the tragic story of vade cultured cells and to transit the endocytic pathway for at Mary Mallon (“Typhoid Mary”), an Irish cook who was chronically least 3 h to receive the environmental cues that result in the infected with S. Typhi and inadvertently spread the disease to many stimulation of toxin gene expression (20, 23). These observations households in the New York city area in the early 1900s. Due to the explained why, despite the intense scrutiny of this pathogen for lack of a treatment that would stop her from shedding S. Typhi and more than a century, this toxin activity had remained undetected. her refusal to abandon her much-loved profession, Mary Mallon Toxicity was shown to be associated with a protein encoded went on to spend the rest of her life in confinement (10). Despite its within a genomic islet that exhibits very significant amino acid long-standing presence in human history, typhoid fever is still today sequence similarity to CdtB, the active subunit of CDT (20). Similar a major public health concern claiming the lives of ∼200,000 people, mostly children in developing countries (2, 11). Significance Unlike other Salmonella enterica (“nontyphoidal”) serovars such as S. Typhimurium, which can infect a broad range of an- Salmonella Typhi is the cause of typhoid fever, a disease that imal hosts generally causing self-limiting gastroenteritis (12, 13), has challenged humans throughout history but that continues S. Typhi is an exclusive human pathogen, where it causes severe to be a major public health concern resulting in ∼200,000 systemic illness (1, 3, 4, 14). The genome sequence of S. Typhi has deaths every year. The recent discovery of typhoid toxin has provided major insight into its evolutionary history, and phyloge- provided unique opportunities to develop much needed pre- netic analysis indicates that this bacterium is highly monomorphic ventive and therapeutic strategies. and may have entered the human population relatively recently (15– 17). The S. Typhi genome harbors a higher-than-expected number Author contributions: J.E.G. wrote the paper. of pseudogenes, an indication that the process of human–host ad- Reviewers: G.D., Wellcome Trust Sanger Institute; and S.F., Stanford University. aptation is resulting in the reduction of its genome. Traditionally, The author declares no conflict of interest. the mechanisms of S. Typhi pathogenesis have been inferred from 1Email: [email protected].

www.pnas.org/cgi/doi/10.1073/pnas.1606335113 PNAS Early Edition | 1of7 Downloaded by guest on September 27, 2021 cytolethal distending toxin. The linear arrangement of the typhoid toxin complex determines that there are no interactions between CdtB and PltB, which explains why deleting pltA results in complete loss of CdtB-mediated toxicity (23). The association of PltA with the PltB oligomer is largely mediated by a short helix at the carboxyl terminal end of PltA, which inserts into the hydrophobic lumen of the PltB channel. In contrast, a disulfide bond between Cys214 of PltA and Cys269 of CdtB covalently links these two subunits (Fig. 2). The structure indicates that there is minimal interaction between CdtB and PltA other than the disulfide bond, which is corroborated by the observation that the complex disassembles upon reduction of this bond (24). Interestingly, despite the high amino acid sequence conservation between PltB and CdtB and its homologs in other toxins, the Cys residue that tethers these two subunits together is unique to the typhoid toxin subunits. Therefore, the evolution of uniquely positioned Cys residues in both PltA and CdtB allowed the tethering of CdtB to the PltA/PltB complex, thus giving rise to a powerful new toxin from the merger of components of other bacterial toxins. Typhoid Toxin Secretion and Export from Mammalian Cells Once synthesized by intracellular bacteria, typhoid toxin is secreted from S. Typhi into the lumen of the bacteria-containing vacuole by a specific and unique protein secretion system (29) (Fig. 3). Although many mechanistic details of this system have yet to be uncovered, it seems that the secretion mechanism is a recent exaptation of the holin/endolysin system used by bacteriophages to exit from infected bacterial cells (30). A central Fig. 1. Typhoid toxin induces cell cycle arrest in intoxicated cells. A diagram of the genetic locus that encodes the typhoid toxin genes is depicted (Top). component of the typhoid toxin secretion system is TtsA (for Culture epithelial cells were infected with WT S. Typhi or a typhoid toxin- typhoid toxin secretion protein A), which is encoded immedi- deficient mutant derivative (ΔcdtB), and, 72 h after infection, the infected ately adjacent to the typhoid toxin genes (29). TtsA is a putative cells were examined by phase contrast microscopy (Middle) or processed to N-acetyl-β-D-muramidase that belongs to a distinct class of bac- measure DNA content by flow cytometry (Bottom). The peaks corresponding teriophage endolysins. During phage release, endolysins are to cells in G0-G1, S, or G2 are indicated. transported through the inner membrane by holins, a family of small membrane proteins that assemble into a protein channel through which endolysins reach the periplasmic space and trigger to CDT, the morphological changes and cell cycle arrest observed in bacterial lysis (31). Holins are most often encoded in the vicinity S. Typhi-intoxicated cells were shown to be due to DNA damage of their cognate endolysins. However, no proteins with features caused by the DNase activity of this CdtB homolog (21). Within the similar to those expected of a functional holin are encoded in the same genomic islet encoding cdtB, ORFs encoding homologs of the vicinity of TtsA, suggesting that it may function in conjunction with “ ” ADP ribosyl transferase A subunit and one of the components of a holin encoded elsewhere in the S. Typhi chromosome. Indeed, “ ” the heteromeric B subunit of pertussis toxin (thus named pltA and cross-talk among different endolysins and holins has been previously pltB, respectively, for pertussis-like toxin A and B) were also iden- reported (30). It is expected that the delivery and/or activity of TtsA tified (23). Unexpectedly, mutations in either pltA or pltB resulted in must be locally restricted because the TtsA-dependent secretion of total loss of the cdtB-dependent toxic activity functionally linking these typhoid toxin is not accompanied by bacterial lysis (29). Consistent three genes. Additional biochemical experiments went on to demon- with this hypothesis, a detailed mutagenesis and domain-swapping strate that CdtB, PltA, and PltB form a complex, an observation that analysis of TtsA identified specificity determinants for typhoid formalized the discovery of typhoid toxin (23). Consequently, toxin-secretion functions that are located within its peptidoglycan- typhoid toxin seemed to have evolved from the combination of binding domain (29). In fact, a phage endolysin could be rendered the activities of two exotoxin ancestors, CDT and pertussis toxins. suitable for functions related to typhoid toxin secretion by in- troducing just a limited number of amino acid changes within its Insights from the Atomic Structure of Typhoid Toxin: How peptidoglycan binding. These observations suggest that the exap- Two Toxins Became One tation of the phage endolysin to perform protein secretion functions The crystal structure of typhoid toxin showed a pyramid-shaped must be a recent evolutionary event. Recently, the secretion of complex composed of five PltB molecules (located at the base) another toxin was shown to occur through a similar mechanism and one molecule each of PltA and CdtB (located at the vertex (32), indicating that, as originally proposed (29), the typhoid toxin “ ” of the pyramid) (24) (Fig. 2). The A2B5 organization of two A secretion system defines a novel protein secretion mechanism used subunits linked to the same “B” subunit observed in typhoid by bacteria to secrete toxins or large extracellular enzymes. toxin is unprecedented among other known members of the AB5 After its secretion into the lumen of the Salmonella-containing toxin family, such as cholera, shiga, or pertussis toxins, which vacuole, typhoid toxin is packaged into vesicle-carrier intermediates, have only one A subunit (25, 26). The PltB oligomer is arranged and it is subsequently transported to the extracellular space where it as a pentamer, with a central channel that is lined by five helixes, can reach its target cells (Fig. 3) (23). The typhoid toxin export and exhibits structural similarity to B subunits of other AB toxins pathway is incompletely understood, but it requires the activity of (27, 28). The PltB protomer shows a typical oligosaccharide- the Rab GTPases Rab29 and Rab31 (33). Importantly, typhoid binding fold located on the side of the pentamer, a location similar toxin cannot intoxicate the cells in which it is produced without first to that of toxins that preferentially bind glycoproteins (25). As becoming extracellular. Addition of an antibody to the extracellular predicted by the amino acid sequence similarities, the PltA and space, for example, effectively protects S. Typhi-infected cells from CdtB subunits exhibit a very similar structure to the pertussis toxin intoxication, supporting the notion that typhoid toxin can intoxicate S1 (and other ADP ribosyl transferases), and to the CdtB subunit of infected cells only through a paracrine/autocrine pathway (23).

2of7 | www.pnas.org/cgi/doi/10.1073/pnas.1606335113 Galán Downloaded by guest on September 27, 2021 INAUGURAL ARTICLE MICROBIOLOGY

Fig. 2. Atomic structure of typhoid toxin. (A) Two views of the overall structure of the typhoid holotoxin complex shown as a ribbon diagram and related by 90° rotation about a vertical axis. CdtB, PltA, and PltB are shown in blue, red, and green, respectively. (B) Bottom view of the channel formed by the PltB pentamer (in green), depicting the PltA C-terminal α-helix (in red) within it. (C) Atomic interface between CdtB and PltA. (Inset) A detailed view of a critical disulphide bond between PltACys 214 and CdtBCys 269 (adapted from ref. 24).

Consequently, cells lacking typhoid toxin receptors and harboring of circulating leukocytes (24), another pathognomonic symptom of S. Typhi could serve as a toxin source while themselves not being a the acute phase of typhoid fever (34). The development of symp- target of its activity, an aspect of the biology of typhoid toxin that toms was correlated with the catalytic activity of CdtB because a may be central to its potential role during infection (see Typhoid typhoid toxin preparation carrying a catalytic mutant of its CdtB Toxin and Typhoid Fever). subunit was unable to induce any symptomatology (24). In contrast, a typhoid toxin preparation carrying a catalytic mutant of its PltA Typhoid Toxin and Typhoid Fever triggered the same response as WT toxin, indicating that this The unique clinical presentation of typhoid fever contrasts the subunit does not play a role in this activity. These observations, clinical presentation of infections by nontyphoidal Salmonella combined with observations in experimentally infected chimpan- serovars, which most often result in a self-limiting gastroenteritis zees (see Typhoid Toxin and S. Typhi Host Specificity), suggest a (e.g., “food poisoning”) (12–14). The molecular bases for the direct role for typhoid toxin in the development of the acute phase differences in these clinical presentations have been a long- symptoms of typhoid fever. It is also likely that typhoid toxin may standing mystery in the field. Although the unique clinical pre- play additional roles during S. Typhi infection. For example, its sentation associated with typhoidal Salmonellae infections are ability to cause leukopenia indicates that typhoid toxin may con- likely to be multifactorial, typhoid toxin is likely to play a central tribute to the infectivity of S. Typhi although human volunteer role in the development of symptoms specific to typhoid fever. studies will be required to investigate this possibility. Finally, it is This hypothesis is supported by the observation that typhoid possible that typhoid toxin’s main role is exerted during S. Typhi toxin is encoded by both typhoidal Salmonella serovars S. Typhi persistent infection. In fact, experiments in humanized mice sug- and S. Paratyphi whereas it is largely absent from nontyphoidal gest this possibility (35). S. Typhi is known to establish chronic, life- serovars. More importantly, administration of purified typhoid long infections in the gall bladder of convalescent individuals, likely toxin into experimental animals was able to reproduce many of a central attribute in its evolutionary success as a host-specific the pathognomonic symptoms of the acute phase of the disease pathogen with no known reservoir. The general mechanisms that (24). Inoculated animals showed signs of stupor and malaise, lead to the carrier state are very poorly understood, and the spe- commonly observed during the acute phase of typhoid fever (1). cific mechanisms by which S. Typhi avoids its clearance by the However, these animals did not develop fever, suggesting that strong immune response mounted by persistent carriers are un- some of the typhoid fever-associated symptoms may not just be a known. Given the well-known toxic effect of CdtB homologs on consequence of the inflammatory response to infection, but, immune cells (36, 37), it is possible that typhoid toxin exported rather, they may be the result of the involvement of the central from infected cells to the immediately surrounding space may help nervous system in the pathogenesis of typhoid fever. Animals to create a local environment in which this pathogen can avoid its that received active toxin also showed a markedly reduced number detection by the immune system.

Galán PNAS Early Edition | 3of7 Downloaded by guest on September 27, 2021 Fig. 3. Model for typhoid toxin secretion and export. Typhoid toxin expression is stimulated after S. Typhi enters into host cells, where it receives the specific cues to initiate its synthesis. The toxin subunits are secreted into the bacteria periplasm (Inset) via the sec machinery, which recognizes canonical secretion signals (secSP) in each one of the polypeptides. After assembly within the bacterial periplasm, the typhoid holotoxin is translocated through the peptidoglycan layer with the help of TtsA, an N-acetyl-β-D-muramidase that shares amino acid sequence similarity to a novel class of bacteriophage endolysins. Like its homologs TtsA lacks a typical secretion signal therefore it is expected that it reaches the periplasm aided by an as yet unidentified holin, a family of small membrane proteins that share the property of forming a protein channel through which endolysins reach the periplasmic space. After secretion from the bacterial cell into the Salmonella-containing vacuole, typhoid toxin is packaged into vesicle carrier intermediates, which transport the toxin to the plasma membrane for its delivery to the extracellular space. The toxin can then reach its cellular targets via paracrine or autocrine pathways.

Typhoid Toxin and S. Typhi Host Specificity with a single polypeptide, PltB, which forms its homomeric B subunit The remarkable host specificity for humans exhibited by S. Typhi (24, 43). Remarkably, however, typhoid toxin does not bind to oth- and S. Paratyphi are incompletely understood and are most likely erwise identical glycans terminated in Neu5Gc (i.e., differing by a multifactorial. A major factor is clearly its inability to infect non- single oxygen atom) (43). This observation provided major insight human hosts (see Salmonella Typhi Host Restriction: Insights from into the molecular bases for S. Typhi human-host specificity because the Study of Typhoid Toxin). However, chimpanzees can be effi- sialoglycans on human cells are unusual in that they are primarily ciently infected with S. Typhi and yet they do not develop typhoid terminated in N-acetylneuraminic acid (Neu5Ac) (44). In other pri- fever(38,39).Infact,S. Typhi reaches significantly higher numbers mates (e.g., chimpanzees) and most other mammals, glycans are in the blood stream of experimentally infected chimpanzees than largely terminated in N-glycolylneuraminic acid(Neu5Gc).Therea- those observed in infected humans. Despite efficient infection, it has son for this difference is the presence of a mutation in the human been previously observed that the symptoms in the experimentally gene that encodes the enzyme CMP-N-acetylneuraminic acid infected chimpanzee are significantly different from those observed hydroxylase (CMAH), which converts Neu5Ac to Neu5Gc. This in humans. For example, one study described the clinical course of mutation is thought to have emerged after hominids separated the disease in chimpanzees as “mild and brief” with “the hyper- from other primates. Consistent with its inability to bind Neu5Gc- toxicity, typhoid facies, stupor, extreme lethargy, so generally as- terminated glycans, typhoid toxin does not bind to chimpanzee sociated with the disease in man, not discernible in the infected cells and tissues, thus explaining why, despite significant replication, chimpanzees” (38). In other words, the S. Typhi-infected animals S. Typhi does not cause typhoid fever in chimpanzees (38, 39). exhibited symptoms more in line with those of a relatively mild Mice, like most mammals, express a functional CMAH. However, “nontyphoidal Salmonella” infection than those of typhoid fever. in the mouse, the expression pattern of CMAH is variable, and The identification of typhoid toxin’s host receptors has provided an therefore mouse tissues display sialoglycans terminated in both explanation for the discrepancies between the clinical presentations Neu5Ac and Neu5Gc (45). This observation explains why, in the of S. Typhi infection in humans and chimpanzees and has led to mouse, typhoid toxin is capable of inducing typhoid fever symptoms major insights into the molecular bases of S. Typhi host specificity. (24). However, a mouse engineered to constitutively express the In epithelial cells, the typhoid toxin receptor is podocalyxin-like CMAH gene and thus displaying only Neu5Gc-terminated glycans protein 1 (PODXL) (24), a polarly localized member of the CD34 was shown to be totally refractory to typhoid toxin (43). Therefore, sialomucin protein family that is also expressed in vascular endo- the exquisite binding selectivity for glycans that are predominantly thelial cells (40, 41). In contrast, in macrophages and T and B cells, expressed in human cells provides an explanation for the inability the receptor is CD45 (24), which is ubiquitously expressed in all of S. Typhi to cause typhoid fever in nonpermissive species that, hematopoietic cells other than erythrocytes and platelets (42). Ty- like chimpanzees, allow significant bacterial replication. These phoid toxin recognizes common glycan moieties on these heavily observations also further support the role of typhoid toxin as a glycosylated surface proteins (24). A detailed glycoarray analysis de- central factor in the development of clinical typhoid fever. termined that typhoid toxin binds a diverse group of sialylated glycans The atomic structure of PltB bound to its glycan receptor with preferential binding to termini with the consensus sequence provided major insight into the structural bases for typhoid Neu5Ac2-3Galβ1-3/β1-4Glc/GlcNAc (24, 43). This broad glycan- toxin’s remarkable binding specificity (43). The structure showed binding specificity resembles the binding properties of pertussis toxin. the canonical glycan-binding site on the side of the PltB pentameric However, in contrast to pertussis toxin, which broadens its binding ring making contact with Neu5Ac through multiple direct and water- specificity through heterogeneity in each of its four heteromeric B- mediated hydrogen bonds with Tyr33, Ser35, Lys59, Thr65, and subunit components, typhoid toxin achieves broad binding specificity Arg100, and hydrophobic contacts with the aromatic rings of Tyr33

4of7 | www.pnas.org/cgi/doi/10.1073/pnas.1606335113 Galán Downloaded by guest on September 27, 2021 and Tyr34 (Fig. 4). The oligosaccharide-binding fold of PltB shares production within infected cells to the extracellular space un-

significant structural similarities with SubB, the B subunits of sub- expectedly led to major insight into the mechanisms that restrict INAUGURAL ARTICLE tilase toxin (24, 27). However, SubB exhibits the opposite specificity, S. Typhi replication in nonhuman hosts (33, 49, 50). By exten- strongly favoring binding to Neu5Gc-terminated glycans (27). sion, these studies shed light on mechanisms used by broad host Comparison of the atomic structures of PltB and SubB bound to Salmonellae to overcome this restriction and thus be able to explore their glycan receptors provided important insight into the structural multiple hosts. Studies aimed at identifying Rab GTPases poten- bases for typhoid toxin’s binding specificity (Fig. 4). The arrange- tially involved in toxin transport led to the serendipitous identifi- ment of the main chain of Neu5Ac relative to the binding pocket of cation of three highly related Rab GTPases (Rab29, Rab32, and PltB is very similar to that of Neu5Gc bound to SubB (27), and many Rab380) that are efficiently recruited to S. Typhi-containing vacu- of the critical interactions between the glycans and specific residues oles but are not recruited to vacuoles containing S. Typhimurium of PltB and SubB are conserved. However, a residue equivalent to (33, 49). These studies also led to the identification of an Tyr78 in SubB, which forms a critical hydrogen bond with the extra S. Typhimurium gene (gtgE) that, when expressed in S. Typhi, was hydroxyl group in Neu5Gc, is missing from PltB (Fig. 4). Instead, at able to prevent the recruitment of the Rab GTPases to the S. Typhi- this position, PltB has the nonpolar residue Val103 and thus is un- containing vacuole. More importantly, expression of gtgE allowed able to interact with Neu5Gc. These observations not only provide a S. Typhi replication in mouse macrophages and mouse tissues, thus structural explanation for typhoid toxin’s inability to bind Neu5Gc- overcoming to a significant extent the host-restriction barrier that terminated glycans but also suggest an evolutionary pathway by prevents these bacteria from replicating in this nonpermissive host. which this toxin may have restricted its binding to human-specific GtgE turned out to be an effector of one of the S. Typhimurium glycans. type III secretion systems, which are specialized nanomachines that deliver bacterial proteins into mammalian cells to modulate cellular Salmonella Typhi Host Restriction: Insights from the Study functions for the pathogen’s benefit (51, 52). GtgE is a specific of Typhoid Toxin protease for Rab29, Rab32, and Rab38, which explains its ability In contrast to most Salmonella enterica serovars, which can infect to block the recruitment of these GTPases to the Salmonella- a broad range of hosts, S. Typhi and S. Paratyphi can infect only containing vacuole (33, 49). Further studies demonstrated that, of humans (1, 3, 12). The strict human host specificity not only has the three GTPases targeted by GtgE, Rab32 plays the most im- been a major impediment for S. Typhi research, but also has had portant role in the pathogen-restriction pathway because S. Typhi a significant influence in its discovery. Karl Joseph Eberth, a was able to efficiently replicate in cells and tissues of mice deficient pathologist at the University of Würzburg, is credited with having in Rab32 or in its exchange factor BLOC3 (50). Therefore, a MICROBIOLOGY observed for the first time the “typhoid bacilli” in samples from Rab32-dependent host-defense mechanism is what prevents infected individuals (46, 47). However, he was unable to culture S. Typhi replication in hosts other than humans (Fig. 5). Immune the organism because he lacked the necessary expertise, which at cells are generally viewed as the central element in host defense that time was available at only a limited number of laboratories. mechanisms. However, the first line of defense against microbial It was Georg Theodor August Gafky, a Robert Koch disciple at infection, and evolutionarily the oldest, is composed of cell- the University of Berlin, who finally cultured the typhoid bacilli, autonomous mechanisms that protect individual cells against which was named Eberthella typhi (48). However, S. Typhi’s strict pathogen invasion (53, 54). These mechanisms synergize with the host specificity prevented Gafky from fulfilling the postulates immune system to confer whole-body protection against patho- enunciated by his influential mentor, Robert Koch. Gafky tire- gens. Rab32 therefore defines a novel cell-autonomous defense lessly tried infecting a variety of animals, including Java mon- mechanism that restricts S. Typhi replication in nonhuman hosts keys, mice, and guinea pigs, but (predictably) failed to reproduce most likely by orchestrating the delivery of an antimicrobial factor the disease, thus failing to fulfill one of the criteria established by to the S. Typhi-containing vacuole (Fig. 5). It is well-established his mentor to identify the causative agent of an infectious dis- that, in specialized cells, Rab32, in conjunction with other Rab ease. He would go on to die without the certainty that he had GTPases and multiprotein assemblies known as BLOC 1, 2, and indeed cultured the typhoid bacilli, and one can imagine unable 3, coordinates the delivery of specific cargo to lysosome-related to convince his mentor of his most important discovery. compartments such as melanosomes and T-cell granules (55–58). Recent work in our laboratory aimed at characterizing the This cargo includes antimicrobial compounds as well as en- cellular pathway that transports typhoid toxin from its site of zymes required for the synthesis of melanin and its phenyl

Fig. 4. Atomic structure of typhoid toxin B subunit bound to its glycan receptor. (A) The atomic structure of the PltB pentamer in complex with the GalNAcβ1-4(Neu5Acα2-8Neu5Acα2-3)Galβ1-4Glc oligosaccharide is shown as a ribbon diagram with each protomer depicted in a different color. Cyan, blue, and red sticks represent the sugar carbon, nitrogen, and oxygen atoms, respectively. (B) Surface charge distribution of the PltB pentamer structure and sugar- binding pockets. (C) Comparison of the sugar-binding sites of PltB and SubB bound to Neu5Ac and Neu5Gc, respectively. Critical residues that differ between SubB (Tyr78) and PltB (Val103) are highlighted as sticks. Other interacting amino acids and sugars are shown in lines. PltB is shown in yellow, Neu5Ac in Cyan, SubB in Green, and Neu5Gc in light purple (adapted from ref. 43 with permission from Elsevier; www.sciencedirect.com/science/journal/00928674).

Galán PNAS Early Edition | 5of7 Downloaded by guest on September 27, 2021 Fig. 5. Model for the Rab32/BLOC-3–dependent cell-autonomous defense pathway that restricts the growth of S. Typhi in nonhuman hosts and the mechanisms evolved by the broad-host S. Typhimurium to counter it. In nonpermissive hosts, Rab32 (and its exchange factor BLOC3) coordinate the delivery of an antimicrobial factor to the Salmonella Typhi-containing vacuole. The broad host S. Typhimurium counters this host-defense pathway by delivering two type III secretion effector proteins, GtgE and SopD2, which exert their function as specific protease and GAP for Rab32, respectively (adapted from ref. 50 with permission from Elsevier; www.sciencedirect.com/science/journal/19313128).

oxidase intermediates, many of which have potent antimi- of novel and sorely needed diagnostic, therapeutic, and prevention crobial activity (56, 58–61). The nature of the antimicrobial strategies to combat typhoid fever. The rather unique distribution of factor or factors that limit Salmonella replication in non- typhoid toxin largely restricted to S. Typhi and S. Paratyphi, coupled specialized cells, however, is still unknown, and, although cells to its strong immunogenicity, suggests that it could serve as the basis other than melanocytes do not make melanine, the machinery that for the development of much needed, cost-effective serological tests orchestrates its synthesis is largely present in most cells. It should that could help not only in the diagnosis of typhoid fever but also in be noted that, in insects, this pathway has been shown to play a the more reliable study of its incidence and epidemiology. The po- central role in the defense response against microbial pathogens tential link of typhoid toxin to the life-threatening symptomatology of (62, 63). In addition to GtgE, broad host Salmonellae such as S. Typhi- the acute phase of typhoid fever raises the possibility of developing murium use a second effector protein, SopD2, that also counters antitoxin therapeutic strategies that, in combination with standard the Rab32-dependent host defense pathway (50). SopD2, how- antimicrobials, could significantly improve the outcome of typhoid ever, does not have protease activity, but it inactivates Rab32 fever treatment. Finally, given the good track record of other bac- functioning as GAP for this GTPase. An S. Typhimurium mutant terial toxoids as immunogens (e.g., tetanus and diphtheria toxoids) strain simultaneously lacking gtgE and sopD2 has a drastic viru- and the demonstrated immunogenicity of typhoid toxin, its in- lence phenotype. C57BL/6 mice receiving ∼1,000 WT LD50 of corporation into vaccine formulations may lead to a much-needed the ΔgtgE ΔsopD2 S. Typhimurium mutant strain were com- effective vaccine not only against typhoid fever caused by S.Typhi pletely refractory to infection, showing no symptoms and rapidly but also against S. Paratyphi A, against which no vaccines are clearing the infection. The virulence defect of the ΔgtgE ΔsopD2 currently available. It is yet unclear whether typhoid toxin con- S. Typhimurium mutant strain was completely reversed in C57BL/6 tributes to the infectivity of S. Typhi, and human volunteer studies mice deficient for BLOC3 or Rab32, demonstrating that these two will be necessary to address this issue. However, the linkage of effector proteins specifically target this host defense pathway (50). typhoid toxin to the development of typhoid fever symptoms Taken together, these observations highlight the importance of the suggests that, at the very least, a strong antitoxin immunity should Rab32-dependent host defense pathway in the control of an intracellular vacuolar pathogen. Furthermore, because both gtgE and be able to protect against the disease. sopD2 are either absent (gtgE)orapseudogene(sopD2)inS. Typhi Concluding Remarks (15), these observations also explain the inability of this human- specific pathogen to explore other hosts. The absence of these ef- For over a century, scientists have wondered why S. Typhi and fectors from S. Typhi would suggest that the Rab32-dependent S. Paratyphi infections had such unique symptomatology and restriction pathway is not functional in humans. However, the pathogenesis. They have also wondered why these pathogens can components of this pathogen-restriction pathway are highly con- cause disease only in humans. The discovery of typhoid toxin has served, making this hypothesis highly unlikely. Instead, these find- provided major insight into both of these fundamental questions. ings suggest that, in humans, S. Typhi must be located within cells Furthermore, it may have provided the long sought “Achilles different from those occupied by S. Typhimurium in mice and that, heel” of Salmonella Typhi, which might lead to the eventual erad- in those cells, this pathway may not be operational. In fact, a recent ication of typhoid fever. Studies that started as a detour from the genome-association study has identified a polymorphism within main focus of my laboratory have provided exciting and productive Rab32 that correlates with increased susceptibility to the human- years of research. I hope that these discoveries will one day fulfill specific pathogen Mycobacterium leprae (64), suggesting that this their promise and contribute to the eradication of typhoid fever. pathway is operational in humans and is also important for the control of other intracellular vacuolar pathogens. ACKNOWLEDGMENTS. I thank present and past colleagues who have contributed to the work in my laboratory that is discussed in this article, Opportunities for Novel Therapeutic and Prevention Maria Lara-Tejero for helpful discussion and the preparation of the figures, Strategies and current members of the J.E.G. laboratory for critical reading of this manuscript. Work in my laboratory relevant to this article is supported by The discovery of typhoid toxin and its role in the pathogenesis of National Institute of Allergy and Infectious Diseases Grants AI055472 typhoid fever has provided unique opportunities for the development and AI079022.

6of7 | www.pnas.org/cgi/doi/10.1073/pnas.1606335113 Galán Downloaded by guest on September 27, 2021 1. Parry CM, Hien TT, Dougan G, White NJ, Farrar JJ (2002) Typhoid fever. N Engl J Med 34. Connor BA, Schwartz E (2005) Typhoid and paratyphoid fever in travellers. Lancet 347(22):1770–1782. Infect Dis 5(10):623–628. INAUGURAL ARTICLE 2. Crump JA, Mintz ED (2010) Global trends in typhoid and paratyphoid Fever. Clin Infect 35. Song J, et al. (2010) A mouse model for the human pathogen Salmonella typhi. Cell Dis 50(2):241–246. Host Microbe 8(4):369–376. 3. Raffatellu M, Wilson RP, Winter SE, Bäumler AJ (2008) Clinical pathogenesis of ty- 36. Shenker BJ, et al. (1999) Actinobacillus actinomycetemcomitans immunosuppressive phoid fever. J Infect Dev Ctries 2(4):260–266. protein is a member of the family of cytolethal distending toxins capable of causing a 4. Wain J, Hendriksen RS, Mikoleit ML, Keddy KH, Ochiai RL (2015) Typhoid fever. Lancet G2 arrest in human T cells. J Immunol 162(8):4773–4780. – 385(9973):1136 1145. 37. Svensson LA, Tarkowski A, Thelestam M, Lagergård T (2001) The impact of Haemo- 5. Dougan G, Baker S (2014) Salmonella enterica serovar Typhi and the pathogenesis of philus ducreyi cytolethal distending toxin on cells involved in immune response. – typhoid fever. Annu Rev Microbiol 68:317 336. Microb Pathog 30(3):157–166. 6. Baker S, Karkey A, Parry C (2014) Are we adequately prepared for the emergence of 38. Edsall G, et al. (1960) Studies on infection and immunity in experimental typhoid Salmonella enterica serovar Paratyphi A? Lancet Glob Health 2(4):e195–e196. fever. I. Typhoid fever in chimpanzees orally infected with Salmonella typhosa. J Exp 7. Fangtham M, Wilde H (2008) Emergence of Salmonella paratyphi A as a major cause Med 112:143–166. of enteric fever: Need for early detection, preventive measures, and effective vac- 39. Metchnikoff E, Besredka A (1911) Recherches sur la fievre typhoide experimental. cines. J Travel Med 15(5):344–350. Ann Inst Pasteur (Paris) 25:865–814. 8. Thucydides (1965) The History of the Peloponnesian War (Library of Alexandria, 40. Yu CY, et al. (2007) A bipartite signal regulates the faithful delivery of apical domain Athens, Greece), revised Ed. marker podocalyxin/Gp135. Mol Biol Cell 18(5):1710–1722. 9. Papagrigorakis MJ, Yapijakis C, Synodinos PN, Baziotopoulou-Valavani E (2006) DNA 41. Furness SG, McNagny K (2006) Beyond mere markers: Functions for CD34 family of examination of ancient dental pulp incriminates typhoid fever as a probable cause of sialomucins in hematopoiesis. Immunol Res 34(1):13–32. the Plague of Athens. Int J Infect Dis 10(3):206–214. 42. Hermiston ML, Zikherman J, Zhu JW (2009) CD45, CD148, and Lyp/Pep: Critical 10. Bourdain A (2010) Typhoid Mary (Bloomsbury, New York). E-book Ed. 11. Buckle GC, Walker CL, Black RE (2012) Typhoid fever and paratyphoid fever: Sys- phosphatases regulating Src family kinase signaling networks in immune cells. – tematic review to estimate global morbidity and mortality for 2010. J Glob Health Immunol Rev 228(1):288 311. 2(1):010401. 43. Deng L, et al. (2014) Host adaptation of a bacterial toxin from the human pathogen – 12. Ohl ME, Miller SI (2001) Salmonella: A model for bacterial pathogenesis. Annu Rev Salmonella Typhi. Cell 159(6):1290 1299. Med 52:259–274. 44. Varki NM, Strobert E, Dick EJ, Jr, Benirschke K, Varki A (2011) Biomedical differences 13. Grassl GA, Finlay BB (2008) Pathogenesis of enteric Salmonella infections. Curr Opin between human and nonhuman hominids: Potential roles for uniquely human as- Gastroenterol 24(1):22–26. pects of sialic acid biology. Annu Rev Pathol 6:365–393. 14. House D, Bishop A, Parry C, Dougan G, Wain J (2001) Typhoid fever: Pathogenesis and 45. Hedlund M, et al. (2007) N-glycolylneuraminic acid deficiency in mice: Implications for disease. Curr Opin Infect Dis 14(5):573–578. human biology and evolution. Mol Cell Biol 27(12):4340–4346. 15. Parkhill J, et al. (2001) Complete genome sequence of a multiple drug resistant Sal- 46. Keating J (1892) The etiology of typhoid fever. Physician Surg 14:469–472. monella enterica serovar Typhi CT18. Nature 413(6858):848–852. 47. Anonymous (1927) Obituary: Professor Karl Joseph Eberth. Br Med J 1(3443):44–45. 16. Baker S, Dougan G (2007) The genome of Salmonella enterica serovar Typhi. Clin 48. Robinson G (1970) Gaffky, Gerog Theodor August. Dicitonary of Scientific Biographies,ed Infect Dis 45(Suppl 1):S29–S33. Gillispie CC (Scribners, New York).

17. Achtman M (2008) Evolution, population structure, and phylogeography of geneti- 49. Spanò S, Galán JE (2012) A Rab32-dependent pathway contributes to Salmonella MICROBIOLOGY cally monomorphic bacterial pathogens. Annu Rev Microbiol 62:53–70. typhi host restriction. Science 338(6109):960–963. 18. Sabbagh SC, Forest CG, Lepage C, Leclerc JM, Daigle F (2010) So similar, yet so dif- 50. Spanò S, Gao X, Hannemann S, Lara-Tejero M, Galán JE (2016) A bacterial pathogen ferent: Uncovering distinctive features in the genomes of Salmonella enterica sero- targets a host Rab-family GTPase defense pathway with a GAP. Cell Host Microbe vars Typhimurium and Typhi. FEMS Microbiol Lett 305(1):1–13. 19(2):216–226. 19. McClelland M, et al. (2001) Complete genome sequence of Salmonella enterica se- 51. Galán JE, Lara-Tejero M, Marlovits TC, Wagner S (2014) Bacterial type III secretion rovar Typhimurium LT2. Nature 413(6858):852–856. systems: Specialized nanomachines for protein delivery into target cells. Annu Rev 20. Haghjoo E, Galán JE (2004) Salmonella typhi encodes a functional cytolethal dis- Microbiol 68:415–438. tending toxin that is delivered into host cells by a bacterial-internalization pathway. 52. Galán JE (2001) Salmonella interactions with host cells: Type III secretion at work. – Proc Natl Acad Sci USA 101(13):4614 4619. Annu Rev Cell Dev Biol 17:53–86. 21. Lara-Tejero M, Galán JE (2000) A bacterial toxin that controls cell cycle progression as 53. Randow F, MacMicking JD, James LC (2013) Cellular self-defense: How cell-autono- – a deoxyribonuclease I-like protein. Science 290(5490):354 357. mous immunity protects against pathogens. Science 340(6133):701–706. 22. Lara-Tejero M, Galán JE (2002) Cytolethal distending toxin: Limited damage as a 54. Deretic V (2011) Autophagy in immunity and cell-autonomous defense against in- strategy to modulate cellular functions. Trends Microbiol 10(3):147–152. tracellular microbes. Immunol Rev 240(1):92–104. 23. Spanò S, Ugalde JE, Galán JE (2008) Delivery of a Salmonella Typhi exotoxin from a 55. Raposo G, Marks MS (2007) Melanosomes: Dark organelles enlighten endosomal host intracellular compartment. Cell Host Microbe 3(1):30–38. membrane transport. Nat Rev Mol Cell Biol 8(10):786–797. 24. Song J, Gao X, Galán JE (2013) Structure and function of the Salmonella Typhi chi- 56. Dell’Angelica EC (2004) The building BLOC(k)s of lysosomes and related organelles. maeric A(2)B(5) typhoid toxin. Nature 499(7458):350–354. Curr Opin Cell Biol 16(4):458–464. 25. Beddoe T, Paton AW, Le Nours J, Rossjohn J, Paton JC (2010) Structure, biological 57. Luzio J, Hackmann Y, Dieckmann N, Griffiths G (2014) The biogenesis of lysosomes functions and applications of the AB5 toxins. Trends Biochem Sci 35(7):411–418. and lysosome-related organelles. Cold Spring Harb Perspect Biol 6:a016840. 26. Merritt EA, Hol WG (1995) AB5 toxins. Curr Opin Struct Biol 5(2):165–171. 58. Bultema JJ, Ambrosio AL, Burek CL, Di Pietro SM (2012) BLOC-2, AP-3, and AP-1 27. Byres E, et al. (2008) Incorporation of a non-human glycan mediates human suscep- proteins function in concert with Rab38 and Rab32 proteins to mediate protein tibility to a bacterial toxin. Nature 456(7222):648–652. – 28. Paton AW, et al. (2006) AB5 subtilase cytotoxin inactivates the endoplasmic reticulum trafficking to lysosome-related organelles. J Biol Chem 287(23):19550 19563. chaperone BiP. Nature 443(7111):548–552. 59. Aspengren S, Hedberg D, Sköld HN, Wallin M (2009) New insights into melanosome – 29. Hodak H, Galán JE (2013) A Salmonella Typhi homologue of bacteriophage mur- transport in vertebrate pigment cells. Int Rev Cell Mol Biol 272:245 302. amidases controls typhoid toxin secretion. EMBO Rep 14(1):95–102. 60. Benado A, Nasagi-Atiya Y, Sagi-Eisenberg R (2009) Protein trafficking in immune cells. – 30. Wang IN, Smith DL, Young R (2000) Holins: The protein clocks of bacteriophage in- Immunobiology 214(7):507 525. fections. Annu Rev Microbiol 54:799–825. 61. Denat L, Kadekaro AL, Marrot L, Leachman SA, Abdel-Malek ZA (2014) Melanocytes 31. Young R (2002) Bacteriophage holins: Deadly diversity. J Mol Microbiol Biotechnol as instigators and victims of oxidative stress. J Invest Dermatol 134(6):1512–1518. 4(1):21–36. 62. Jiang H, Vilcinskas A, Kanost MR (2010) Immunity in lepidopteran insects. Adv Exp 32. Hamilton JJ, et al. (2014) A holin and an endopeptidase are essential for chitinolytic Med Biol 708:181–204. protein secretion in Serratia marcescens. J Cell Biol 207(5):615–626. 63. Lu A, et al. (2014) Insect prophenoloxidase: The view beyond immunity. Front Physiol. 33. Spanò S, Liu X, Galán JE (2011) Proteolytic targeting of Rab29 by an effector protein 5:252. distinguishes the intracellular compartments of human-adapted and broad-host 64. Zhang F, et al. (2011) Identification of two new loci at IL23R and RAB32 that influence Salmonella. Proc Natl Acad Sci USA 108(45):18418–18423. susceptibility to leprosy. Nat Genet 43(12):1247–1251.

Galán PNAS Early Edition | 7of7 Downloaded by guest on September 27, 2021